Comparison of the Effectiveness of Single-Component and Multicomponent Interventions for Reducing Radiation Doses in Patients Undergoing Computed Tomography: A Randomized Clinical Trial

Abstract

Importance: Computed tomography (CT) radiation doses vary across institutions and are often higher than needed.

Objective: To assess the effectiveness of 2 interventions to reduce radiation doses in patients undergoing CT.

Design, setting, and participants: This randomized clinical trial included 864 080 adults older than 18 years who underwent CT of the abdomen, chest, combined abdomen and chest, or head at 100 facilities in 6 countries from November 1, 2015, to September 21, 2017. Data analysis was performed from October 4, 2017, to December 14, 2018.

Interventions: Imaging facilities received audit feedback alone comparing radiation-dose metrics with those of other facilities followed by the multicomponent intervention, including audit feedback with targeted suggestions, a 7-week quality improvement collaborative, and best-practice sharing. Facilities were randomly allocated to the time crossing from usual care to the intervention.

Main outcomes and measures: Primary outcomes were the proportion of high-dose CT scans and mean effective dose at the facility level. Secondary outcomes were organ doses. Outcomes after interventions were compared with those before interventions using hierarchical generalized linear models adjusting for temporal trends and patient characteristics.

Results: Across 100 facilities, 864 080 adults underwent 1 156 657 CT scans. The multicomponent intervention significantly reduced proportions of high-dose CT scans, measured using effective dose. Absolute changes in proportions of high-dose scans were 1.1% to 7.9%, with percentage reductions in the proportion of high-dose scans of 4% to 30% (abdomen: odds ratio [OR], 0.82; 95% CI, 0.77-0.88; P < .001; chest: OR, 0.92; 95% CI, 0.86-0.99; P = .03; combined abdomen and chest: OR, 0.49; 95% CI, 0.41-0.59; P < .001; and head: OR, 0.71; 95% CI, 0.66-0.76; P < .001). Reductions in the proportions of high-dose scans were greater when measured using organ doses. The absolute reduction in the proportion of high-dose scans was 6.0% to 17.2%, reflecting 23% to 58% reductions in the proportions of high-dose scans across anatomical areas. Mean effective doses were significantly reduced after multicomponent intervention for abdomen (6% reduction, P < .001), chest (4%, P < .001), and chest and abdomen (14%, P < .001) CT scans. Larger reductions in mean organ doses were 8% to 43% across anatomical areas. Audit feedback alone reduced the proportions of high-dose scans and mean dose, but reductions in observed dose were smaller. Radiologist's satisfaction with CT image quality was unchanged and high during all periods.

Conclusions and relevance: For imaging facilities, detailed feedback on CT radiation dose combined with actionable suggestions and quality improvement education significantly reduced doses, particularly organ doses. Effects of audit feedback alone were modest.

Trial registration: ClinicalTrials.gov Identifier: NCT03000751.

Conflict of interest statement

Conflict of Interest Disclosures: Drs Smith-Bindman, López-Solano, and Miglioretti reported receiving grants from the National Institutes of Health (NIH) and the Patient-Centered Outcomes Research Institute (PCORI) during the conduct of the study. Dr Einstein reported receiving grants from Canon Medical Systems USA and Roche Medical Systems and receiving grants and personal fees from W. L. Gore and Associates outside the submitted work. Dr Solberg reported receiving grants from the NIH during the conduct of the study and nonfinancial support from HealthPartners outside the submitted work. Dr Cervantes reported receiving grants from University of California, San Francisco (UCSF) during the conduct of the study. Dr Goode reported receiving grants from the UCSF during the conduct of the study and receiving personal fees from Bayer Healthcare outside the submitted work. Dr Starkey reported receiving personal fees from Bayer Japan outside the submitted work. No other disclosures were reported.

Figures

Figure 1.. CONSORT Flow Diagram

Imaging facility groups were formed by grouping imaging facilities that shared practices and/or personnel to reduce contamination when randomizing. Members of these facility groups were randomized together. Analysis was at the level of the facility. IRB indicates institutional review board.

Figure 2.. Effective Dose Compared With Preaudit Baseline

Effective dose by week before or after the audit, shown as the percentage of the 75th percentile or median dose at pre-audit baseline dose. Numbers on the horizontal cyan lines in A, C, E, and G represent the 75th percentile in dose at baseline and in B, D, F, and H represent the 50th percentile in dose at baseline. Each circle reflects the 75th percentile dose (A, C, E, G) or the median dose (B, D, F, H) relative to the pre-audit baseline values across all imaging facilities. Doses equal to the baseline dose are 100%. Values less than 100% indicate the relative decrease in dose.